Automatic backlight compensation and exposure control

Abstract

In some embodiments, images for the video conference may be white balanced and/or exposure controlled. White balancing may include dividing an image's pixels into multiple ranges according to their luminance and chrominance values. In some embodiments, blue and red analog gains may be modified to bring red and blue accumulations closer to green accumulations. In some embodiments, luminance values for an image may be adjusted by increasing or decreasing exposure. The decision whether to increase or decrease the exposure may be made by looking at statistics for luminance values of skin colored regions if there are sufficient skin colored regions on the image. Otherwise, luminance values for the entire image may be analyzed to determine whether to increase or decrease exposure.

Description

PRIORITY

This application claims priority to U.S. Provisional Patent Application Ser. No. 60/619,210 titled “Video Conference Call System”, which was filed Oct. 15, 2004, whose inventors are Michael J. Burkett, Ashish Goyal, Michael V. Jenkins, Michael L. Kenoyer, Craig B. Malloy, and Jonathan W. Tracey.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates generally to video and, more specifically, to image enhancement.

2. Description of the Related Art

Images in video conference calls may suffer from exposure and/or white balance problems. Exposure problems may be caused by too much or too little light being received through the camera. White balance problems may be caused by incorrect ratios of colors in the image. Correct exposure may be more difficult to achieve than white balance, because white balance may be easier to fix than exposure after the image is acquired.

Gray world algorithms and specular reflection algorithms are two categories of white balance algorithms. Gray world algorithms may look for approximately the same amount of color in Red Green Blue (RGB) space (or equal Cb and Cr in YCC space) and adjust for any offsets. In YCC space, Y is a measure of luminance (intensity of light per unit area), Cb is a measure of blue chrominance (luminance minus blue), and Cr is a measure of red chrominance (luminance minus red). Specular reflection models may assume that the brightest spots in an image reflect the light source neutrally and may adjust the spectrum of the image accordingly. Both gray world algorithms and specular reflection models have disadvantages that may result in poor image quality.

SUMMARY OF THE INVENTION

In some embodiments, images for the video conference may be white balanced and/or exposure controlled. In some embodiments, pixels of interest may be used for determining a proper exposure correction and/or white balancing an image. Pixels of interest may include pixels identified with movement in the image, identified with skin portions in the image, identified with portions of the image closest to the camera, identified with pixels corresponding to a speaking participant, and/or identified with warm portions of the image. In some embodiments, the pixels for the image may be adjusted based on the exposure correction determined for the pixels of interest.

In some embodiments, the exposure correction may be determined by calculating a skew and a predicted peak, and then modifying the exposure based on the calculated predicted peak. In some embodiments, a skew of the luminance values of at least a portion of the image may be calculated, and if the skew is above a predetermined value, a predicted peak value may be calculated according to a first formula. If the skew is not above a predetermined value, the predicted peak value may be calculated according to a second formula. In some embodiments, if the predicted peak value is less than a predetermined low predicted peak value, an exposure of the image may be increased, and if the predicted peak value is more than a predetermined high predicted peak value, the exposure of the image may be decreased.

White balancing may include dividing an image's pixels into multiple ranges according to their luminance and chrominance values. In some embodiments, a video image may be enhanced by modifying blue and red analog gains to bring red and blue accumulations closer to green accumulations. In some embodiments, luminance values for an image may be adjusted by increasing or decreasing exposure. The decision whether to increase or decrease the exposure may be made by looking at statistics for luminance values of skin colored regions if there are sufficient skin colored regions on the image. Otherwise, luminance values for the entire image may be analyzed to determine whether to increase or decrease exposure.

BRIEF DESCRIPTION OF THE DRAWINGS

A better understanding of the present invention may be obtained when the following detailed description is considered in conjunction with the following drawings, in which:

FIG. 1 illustrates a video conference system, according to an embodiment;

FIG. 2 illustrates a method for exposure control for an image in a videoconference, according to an embodiment;

FIG. 3 illustrates a method for exposure control for an image, according to an embodiment;

FIG. 4 illustrates a method for evaluating the exposure of an image using statistics of the Y values for the image, according to an embodiment;

FIG. 5 illustrates a method for increasing or decreasing exposure, according to an embodiment; and

FIG. 6 illustrates a method for accumulating pixels in multiple accumulators, according to an embodiment.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the drawings and detailed description thereto are not intended to limit the invention to the particular form disclosed, but on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present invention as defined by the appended claims. Note, the headings are for organizational purposes only and are not meant to be used to limit or interpret the description or claims. Furthermore, note that the word “may” is used throughout this application in a permissive sense (i.e., having the potential to, being able to), not a mandatory sense (i.e., must). The term “include”, and derivations thereof, mean “including, but not limited to”. The term “coupled” means “directly or indirectly connected”.

DETAILED DESCRIPTION OF THE INVENTION

Incorporation by Reference U.S. Provisional Patent Application Ser. No. 60/619,303, titled “Speakerphone”, which was filed Oct. 15, 2004, whose inventors are Michael L. Kenoyer, William V. Oxford, and Simon Dudley is hereby incorporated by reference in its entirety as though fully and completely set forth herein.

U.S. Provisional Patent Application Ser. No. 60/619,212, titled “Video Conferencing Speakerphone”, which was filed Oct. 15, 2004, whose inventors are Michael L. Kenoyer, Craig B. Malloy, and Wayne E. Mock is hereby incorporated by reference in its entirety as though fully and completely set forth herein.

U.S. Provisional Patent Application Ser. No. 60/619,227, titled “High Definition Camera and Mount”, which was filed Oct. 15, 2004, whose inventors are Michael L. Kenoyer, Patrick D. Vanderwilt, Paul D. Frey, Paul Leslie Howard, Jonathan I. Kaplan, and Branko Lukic, is hereby incorporated by reference in its entirety as though fully and completely set forth herein.

U.S. Provisional Patent Application Ser. No. 60/619,210, titled “Video Conference Call System”, which was filed Oct. 15, 2004, whose inventors are Michael J. Burkett, Ashish Goyal, Michael V. Jenkins, Michael L. Kenoyer, Craig B. Malloy, and Jonathan W. Tracey is hereby incorporated by reference in its entirety as though fully and completely set forth herein.

FIG. 1 illustrates an embodiment of a video conference system. In some embodiments, a camera 103 may capture an image of conference participants (e.g., local participant 107). Images of other conference participants (e.g., from remote conference sites) may be displayed on the display 101. In some embodiments, the local image of the conference participant 107 may need to be white balanced and/or exposure controlled. In some embodiments, images from remote sites may also be enhanced. In some embodiments, image enhancement may be performed by electronics in the camera 103 and/or codec 105. The electronics may include circuitry, processors, memory, etc. for performing image enhancement.

FIG. 2 illustrates a method for exposure control for an image in a videoconference, according to an embodiment. It is noted that in various embodiments one or more of the method elements may be performed concurrently, in a different order, or be omitted. Additional elements may be performed as desired.

At 201, video data may be received for display on a display device in a videoconference. In some embodiments, the video data may include a plurality of pixels corresponding to an image. The image may be a portion of an image in a video stream. In some embodiments, the image may include multiple images and/or multiple portions of an image (e.g., multiple images in a chronological sequence may be considered together). The image may include at least one human participant image.

At 203, a first plurality of pixels corresponding to a first portion of the image may be determined. In some embodiments, the first plurality of pixels may be pixels of interest in the image. In some embodiments, the first plurality of pixels may be determined by determining which pixels of the image to exclude from consideration (e.g., pixels that are not of interest).

In some embodiments, the plurality of pixels may correspond to pixels of a moving object in the image. In some embodiments, algorithms (e.g., implemented in hardware and/or software) may be implemented to determine which pixels are of interest (e.g., correspond to a moving object). For example, in some embodiments, the algorithm may detect pixels corresponding to a moving object by looking for pixels changing luminance values more dramatically and/or more quickly than other pixels of the image.

In some embodiments, the first plurality of pixels may be pixels that correspond to skin portions of the human participant. For example, ranges of pixel values (depending on the color space used) may correlate more closely to human skin than other values. While dark skin and light skin may have different luminance values, in YCC space, they may have similar ranges of blue and red chrominance values.

In some embodiments, the first plurality of pixels may correspond to objects that are closest to the camera. For example, the pixels of an object closest to a camera may be determined using two or more cameras and the parallax effect. In some embodiments, measurement equipment may be used to detect the distance of objects from the camera (e.g., a laser and sensor). These distance measurements may be correlated with the pixels in the image and the first plurality of pixels may be designated as the pixels that correlate to the closest objects.

In some embodiments, a camera position relative to an array of microphones coupled to the videoconference system may be known, and the first plurality of pixels may be determined based on identified portions corresponding to one or more participants in the image who are speaking. The algorithm may correlate the positions of the participants who are speaking (e.g., determined through beamforming) relative to the camera to determine the first plurality of pixels.

In some embodiments, a temperature sensor (e.g., an infrared detector) may be coupled to the camera. The first plurality of pixels may be determined based on which pixels in the image correspond to warm portions of the image (i.e., which may correlate to human bodies that are usually warmer than their surroundings).

In some embodiments, the pixels in a certain portion of an image may be used for the first plurality of pixels. For example, the pixels in the center of the image may be used as the first plurality of pixels.

In some embodiments, other methods for finding pixels of interest may be used. In some embodiments, two or more methods may be used in conjunction. In some embodiments, a method for determining the pixels of interest may be stored and used to determine whether to use the same method on the next frame or whether to switch methods (e.g., the method may be switched if a sufficient number of pixels of interest is not found).

At 205, the first plurality of pixels may be examined to determine an exposure correction for the first plurality of pixels. For example, the luminance values of the pixels of interest may be analyzed to determine if there are luminance values above a predetermined upper limit. If not, the exposure correction may indicate that the exposure of the image needs to be increased. In some embodiments, if there are luminance values above the predetermined upper limit, then the pixels may be analyzed to determine if there are pixels below a predetermined lower limit. If not, the exposure correction may indicate that the exposure for the image needs to be decreased. In some embodiments, an average of the luminance values of the first plurality of pixels may be compared to a threshold to determine an exposure correction. In some embodiments, a histogram of the luminance values of the first plurality of pixels may be used to determine an exposure correction (e.g., the histogram may be compared to a predetermined histogram). The exposure correction may be based on a correction needed to make the current histogram match the predetermined histogram.

At 207, the plurality of pixels may be adjusted corresponding to the image based on the exposure correction. In some embodiments, the entire image may be adjusted based on the determined exposure correction.

FIG. 3 illustrates a method for exposure control for an image, according to an embodiment. It is noted that in various embodiments one or more of the method elements may be performed concurrently, in a different order, or be omitted. Additional elements may be performed as desired.

At 301, video data may be received for display on a display device in a videoconference. In some embodiments, the video data may include a plurality of pixels corresponding to at least one image. In some embodiments, the image may include at least one human participant image.

At 303, a first plurality of pixels corresponding to a first portion of the image may be determined. In some embodiments, the first plurality of pixels may include pixels of interest. In some embodiments, the first plurality of pixels may be based on identified skin portions of at least one human participant image. In some embodiments, each of the plurality of pixels may include a luminance value.

At 305, a skew of the luminance values of at least a portion of the image may be calculated. It is to be understood that the term “skew” as used herein refers to a measurement relative to the third standardized moment of the distribution of the pixel values. For example, the skew of the Y values may be calculated according to:
skew=(average−median)/standard deviation
Where standard deviation is calculated as: $\mathrm{standard}\mathrm{deviation}=\sqrt{\sum _N^i\frac{{\left(x_i-\mathrm{average}\right)}^2}{N-1}}$
where N may be the total number of samples. Other calculations for standard deviation and skew may also be used.

At 307 a determination may be made whether the skew is above a predetermined value.

At 309, if the skew is above a predetermined value, a predicted peak value may be calculated according to a first formula. In some embodiments, according to the first formula, the predicted peak value may be approximately an average of the luminance values minus a standard deviation of the luminance values times a skew of the luminance values (predicted peak value=average−(standard deviation*skew)). In some embodiments, the predicted peak value may not be a realistic luminance value.

At 311, if the skew is not above a predetermined value, the predicted peak value may be calculated according to a second formula. In some embodiments, according to the second formula, the predicted peak value may be approximately a total of the luminance values divided by the number of luminance values.

At 313, a determination may be made whether the predicted peak value is less than a predetermined low predicted peak value.

At 315, if the predicted peak value is less than a predetermined low predicted peak value, an exposure of the image may be increased.

At 317, if the predicted peak value is more than a predetermined high predicted peak value, the exposure of the image may be decreased.

FIG. 4 illustrates a method for evaluating the exposure of an image using statistics of the Y values for the image, according to an embodiment. In various embodiments, the Y values for the entire image or for only pixels of interest (e.g., skin colored pixels) may be evaluated. It is noted that in various embodiments one or more of the method elements may be performed concurrently, in a different order, or be omitted. Additional elements may be performed as desired.

At 401, a percentage of image pixels with red chrominance (Cr) and blue chrominance (Cb) in a predetermined skin range may be determined. In some embodiments, regions of the image with Cb between 75 and 128 and Cr between 130 and 170 may be considered skin colored.

At 403, if the percentage of image pixels Cr and Cb in the predetermined range is above a minimum skin percentage, at 405, a skew of the Y values for the pixels in the predetermined skin range may be calculated. In some embodiments, a minimum percentage of 5 percent may be used. Other percentages may also be used.

At 407, a determination may be made whether the skew is above a predetermined skew minimum. For example, a minimum skew of 0.5 may be used. Other skew minimums may also be used.

At 409, if the skew is above 0.5, the predicted peak value may be calculated according to a first formula (predicted peak value=average−(standard deviation* skew)). In some embodiments, other formulas may be used to calculate the predicted peak value.

At 411, if the skew is not above the skew minimum, an average Y (i.e., a second formula→predicted peak value=total of all Y values/number of Y values) for the pixels in the predetermined range may be used as the predicted peak value.

At 455, if the percentage of image pixels Cr and Cb in the predetermined range is not above a minimum percentage, the skew for the entire image may be calculated.

At 457, a determination may be made whether the skew calculated at 455 is above a predetermined skew minimum. For example, a skew minimum of 0.5 may be used.

At 459, if the calculated skew is above the predetermined skew minimum, the predicted peak value may be calculated according to the first formula:
predicted peak value=average−(standard deviation*skew)

At 461, if the calculated skew is not above the predetermined skew minimum, the average Y value (i.e., the second formula→predicted peak value=total of all Y values/number of Y values) for the entire image may be used as the predicted peak value.

At 413, if the predicted peak value is less than a predetermined low predicted peak value (e.g., 60), at 415, the exposure may be increased. In some embodiments, the image may be adjusted such that the Y values of the skin are spread over a reasonable range.

At 417, if the predicted peak value is greater than a predetermined high predicted peak value (e.g., 200), at 421 the exposure may be decreased.

At 423, the system may wait for the next frame.

In various embodiments, an auto exposure control may be used to adjust the Y values. In some embodiments, the auto exposure control may have two components to evaluate exposure and to modify exposure. Auto exposure control may manipulate several mechanisms to achieve proper exposure. In some embodiments, an aperture stop that controls light passing through the lens may be manipulated by the auto exposure control. For example, the aperture stop may be approximately in the range of f3.5 to f16. The wider the aperture stop is open, the more the depth of field is reduced. Reducing the aperture may increase the depth of field but may also let less light through. In some embodiments, integration time (i.e., how long light is allowed to accumulate on each pixel) may also be manipulated by the auto exposure control. The integration time may be extended as long as possible while achieving the required frame rate. Longer integration times may produce lower noise in the image, but may reduce frame rate. In some embodiments, to achieve 30 frames per second, a maximum desired integration time may be less than 33 milliseconds. In some embodiments, the auto exposure control may also manipulate gain in the amplifiers converting analog voltages into discrete binary values. In some embodiments, the analog gain may be approximately in the range of 1 to 15. However, in some embodiments, increasing the analog gain may amplify noise. The auto exposure control may also manipulate digital gain. Digital gain may also amplify noise.

FIG. 5 illustrates a method for increasing or decreasing exposure, according to an embodiment. It is noted that in various embodiments one or more of the method elements may be performed concurrently, in a different order, or be omitted. Additional elements may be performed as desired.

At 501, the exposure may be evaluated.

At 503, if the exposure needs to be decreased, several possible methods may be selected. In some embodiments, if a method is selected, after the method is applied, the flow may return to 501 where the exposure may again be evaluated. In some embodiments, one of the methods for adjusting exposure may be applied multiple times to the same image. In addition, multiple different methods to adjust exposure may be applied.

At 505, if the integration time is greater than a predetermined minimum integration time (e.g., 33 ms), the integration time may be reduced at 507.

At 511, if digital gain is available, the digital gain may be reduced at 513. The exposure correction may thus be accomplished through pixel gain.

At 517, if analog gain is available, at 519, the analog gain may be reduced.

At 525, if the iris is open, at 527, the aperture of the iris may be reduced.

If none of the above methods are available, or if the exposure still needs to be decreased after one or more of the above methods is applied, the integration time may be further reduced.

At 509, if the exposure needs to be increased, several possible methods may be selected.

At 511, if the integration time is below a predetermined minimum integration time (e.g., 33 ms), the integration time may be increased at 521.

At 523, if the iris is closed, at 529, the iris may be opened.

At 531, if there is analog gain left, at 533, the analog gain may be increased.

At 537, if there is digital gain left, at 539, the digital gain may be increased.

If none of the above methods are available, or if the exposure still needs to be increased after one or more of the above methods is applied, the integration time may be further increased if the integration time is less than a predetermined maximum integration time (e.g., 100 ms).

White Balance

In some embodiments, white balance control may include controlling a range of luminance values and a range of chrominance values for an image to be displayed in a conference call. In some embodiments, a variant of the gray world algorithm and the specular reflection model may be used. This may include identifying a likely light source in an image and correcting the luminance and chrominance values in the image. For example, a light source may include direct sun, clear sky, cloudy sky, incandescent lamps, or other lamps. Other light sources are also possible.

To white balance the image, a correcting algorithm may adjust an image by classifying the pixels of the image into different ranges according to characteristic values of the pixels. For example, if using a YCC color space, the pixels may be classified into different ranges of luminance (Y), blue chrominance (Cb), and red chrominance (Cr). Other color spaces may also be used. Then, depending on the percentage of pixels in each of the different ranges, the image may be corrected accordingly.

FIG. 6 illustrates a method for accumulating pixels in multiple accumulators, according to an embodiment. It is noted that in various embodiments one or more of the method elements may be performed concurrently, in a different order, or be omitted. Additional elements may be performed as desired.

In various embodiments, the pixels may be classified into a bright, brighter, and brightest range. In some embodiments, other numbers of ranges may be used (e.g., four or greater or two or less). In some embodiments, each range may have a Y lower bound (e.g., Y1lo, Y2lo, and Y3lo) and a Y upper bound (e.g., Y1hi, Y2hi, and Y3hi). Each range may also have a Cb upper bound (Cb1hi, Cb2hi, and Cb3hi) and a Cb lower bound (Cb1lo, Cb2lo, and Cb3lo) for Cb, and a Cr upper bound (Cr1hi, Cr2hi, and Cr3hi) and Cr lower bound (Cr1lo, Cr2lo, and Cr3lo) for Cr. The pixels may thus be separated into three ranges (one range with the brightest pixels, then the next brightest pixels, and finally the least brightest pixels). Some pixels may not be in any of the three ranges. In some embodiments, each range may be assigned three accumulators (i.e., in the first range, one for luminance range Y1lo to Y1hi, one for Cb chrominance range Cb1lo to Cb1hi, and one for Cr chrominance range Cr1lo to Cr1hi). There may be a total of nine accumulators (3 more for the second range and three more for the third range). In some embodiments, other numbers of accumulators may be used.

In some embodiments, nine accumulators and three counters may be used to categorize the pixels of an image according to brightness and chrominance. Other numbers of accumulators and counters may also be used. In some embodiments, four register sets may be programmed by software to sort the pixels into the three acceptance ranges (e.g., (Y1, Cb1, Cr2), (Y2, Cb2, Cr2), and (Y3, Cb3, Cr3)). Other numbers of acceptance ranges and register sets may also be used.

In some embodiments, pixels may be converted from one color space to another (e.g., from RGB space to YCC space). At 601, the Y, Cb, and Cr may be compared to the three value ranges. In some embodiments, Y, Cb, and Cr may be provided as arbitrary numbers (e.g., relative to an exposure level of the image). At 603, the Y of the pixel may be compared to the Y3lo to Y3hi range (e.g., 210 to 230). Again, it is to be noted that the numbers 210 and 230 may be arbitrary and may be dependent on the current exposure of the image and the units used for Y, Cb, and Cr (or other color spaces if used). If it has a Y in that range, at 605, the Cb of the pixel may be compared to the Cb3lo to Cb3hi range (e.g., 108 to 148). If the Cb of the pixel is in that range, at 607, the Cr of the pixel may be compared to the Cr3lo to Cr3hi range (e.g., 108 to 148). If the Cr of the pixel is in that range, each of the three accumulators (one for Y3, one for Cb3, and one for Cr3) may store their respective RGB value for the pixel. In addition, at 609, the counter for this set of ranges may be incremented by one.

If the Y of the pixel is not in the Y3lo to Y3hi range, at 611, it is compared to the Y2lo to Y2hi range (e.g., 170 to 210). If it has a Y in that range, at 613, it may be compared to the Cb2lo to Cb2hi range (e.g., 98 to 158). If the Cb of the pixel is in that range, at 615, it may be compared to the Cr2lo to Cr2hi range (e.g., 98 to 158). If the Cr of the pixel is in that range, each of the three accumulators (one for Y3, one for Cb3, and one for Cr3) may store their respective RGB value for the pixel. In addition, at 617, the counter for this set of ranges may be incremented by one.

If the Y of the pixel is not in the Y3lo to Y3hi range, at 621, it is compared to the Y1lo to Y1hi range (e.g., 130 to 170). If it has a Y in that range, at 623, the Cb for the pixel may be compared to the Cb1lo to Cb1hi range (e.g., 88 to 168). If the Cb of the pixel is in that range, at 625, the Cr of the pixel may be compared to the Cr1lo to Cr1hi range (e.g., 88 to 168). If the Cr of the pixel is in that range, each of the three accumulators (one for Y3, one for Cb3, and one for Cr3) may store their respective value for the pixel (which may be an RGB value). In addition, at 627, the counter for this set of ranges may be incremented by one. In some embodiments, pixels with Y below Y1lo or above the Y3hi may be ignored.

In some embodiments, if a sufficient number of pixels (e.g., approximately 0.5%) are in the brightest Y range (e.g., approximately 210 to 230, chrominance between 108 and 148) those accumulations may be used to determine how much to adjust the pixels of the image in order to correct the white balance for the image. In some embodiments, the percentage may be determined using the counter for the range as compared to a total number of pixels for the image. A baseline percentage, e.g., 0.5%, may be predetermined for comparison to the percentage of pixels found in the brightest range. Other percentages may also be used. Because pixels in the brightest range should be the closest to white, the chrominance ranges for Cr and Cb may be relatively small with 128 (neutral) as the center. If there are not at least 0.5% of the pixels in this range, the next range (e.g., Y between 170 and 210, chrominance (for Cr and Cb) between 98 and 158) may be examined. Because this range is not as bright as the first range, a larger range (with 128 in the center) may be used for the chrominance because values in this range may be further from neutral. In some embodiments, more pixels may need to be present in this range (e.g., a baseline percentage of approximately 1%) for these pixels to be used for white balance. Other percentages may also be used for the comparison. In some embodiments, the last range may need approximately 2% (Y between 130 and 170, chrominance between 88 and 168). While examples of percentages have been provided, it is to be understood that the percentages may also be variable (i.e., the percentages may be adjusted based on factors such as previous image exposure).

If the percentages for these ranges are not exceeded, the software may do nothing. In some embodiments, if a percentage for one of the ranges is exceeded, the blue and red gains (e.g., analog and/or digital filters) of sensors for the camera may be modified by a ratio according to a ratio needed to bring red and blue accumulation values closer to green. For example, if sensor values for the green sum is 56711, the red sum is 64379, and the blue sum is 48925, the image may have too much red, and not enough blue (i.e., white may have equal amounts of red, green, and blue). In some embodiments, the red gain may be adjusted by (56711/64379) or 0.8809, and the blue gain may be adjusted by (56711/48925) or 1.159. Other ratios and calculations may also be used to adjust the gains.

In some embodiments, the sensors may be adjusted by other ratios. For example, the ratios may be dependent on which corresponding percentage was exceeded. In some embodiments, the amount of adjustment may be based on the values of Y and chrominance (stored in the accumulations) for the pixels in the accepted range. For example, in some embodiments, weights may be applied based on the percentage of pixels in a particular range and/or based on the actual values of Y and chrominance for pixels in a particular range. Other criteria may also be used. In some embodiments, this characterization of the image may be done when software sets a control bit. In some embodiments, one frame of data may be analyzed at a time. The counters and accumulators may be cleared at the start of each frame when the control is set.

In various embodiments, the distribution of Y values for an image may be evaluated to determine correct exposure. In some embodiments, a histogram of values for the image may be used to determine whether the image has proper exposure. For example, it may be desirable to have an image with Y values above 230 and/or values of below 40.

In various embodiments, the image exposure may need correction if Y values are not in certain ranges. For example, some video standards (e.g. BT656) may only allow certain ranges of Y values (e.g., between 16 (black) and 240 (white)). Other ranges may also be used. In some embodiments, the broader the range of Y values, the better the image. For example, an image with only values between 16 and 40 may be very dark, while an image with values between 200 and 240 may be very white.

In some embodiments, even when the image has Y values in a broad range of histograph bins (e.g., from 16 to 240), it may still need exposure modification. For example, backlighting may require modifying exposure even when the bins have reasonable values. Windows behind people in a conference room may provide more light than the lights in the room. This may cause a highly non-uniform distribution of Y values.

Further modifications and alternative embodiments of various aspects of the invention may be apparent to those skilled in the art in view of this description. Accordingly, this description is to be construed as illustrative only and is for the purpose of teaching those skilled in the art the general manner of carrying out the invention. It is to be understood that the forms of the invention shown and described herein are to be taken as embodiments. Elements and materials may be substituted for those illustrated and described herein, parts and processes may be reversed, and certain features of the invention may be utilized independently, all as would be apparent to one skilled in the art after having the benefit of this description of the invention. Changes may be made in the elements described herein without departing from the spirit and scope of the invention as described in the following claims.

Claims

1. A method for exposure control for an image in a videoconference, comprising:

receiving video data for display on a display device in a videoconference, wherein the video data comprises a plurality of pixels corresponding to at least one image, wherein the at least one image comprises at least one human participant image;

determining a first plurality of pixels corresponding to a first portion of the image;

examining the first plurality of pixels to determine an exposure correction for the first plurality of pixels;

adjusting the plurality of pixels corresponding to at least one image based on the exposure correction.

2. The method of claim 1, wherein said determining the first plurality of pixels is performed based on identified movement in the image.

3. The method of claim 1, wherein said determining the first plurality of pixels is performed based on identified skin portions of the at least one human participant image.

4. The method of claim 1, wherein said determining the first plurality of pixels is performed based on identified portions of the image that are closest to the camera.

5. The method of claim 1, wherein said determining the first plurality of pixels is performed based on identified portions corresponding to one or more participants in the image who are speaking.

6. The method of claim 1, wherein said determining the first plurality of pixels is performed based on identified warm portions of the image.

7. A method for exposure control for an image, comprising:

calculating a skew of the luminance values of at least a portion of the image;

if the skew is above a predetermined value, calculating a predicted peak value according to a first formula;

if the skew is not above a predetermined value, calculating the predicted peak value according to a second formula;

if the predicted peak value is less than a predetermined low predicted peak value, increasing an exposure of the image; and

if the predicted peak value is more than a predetermined high predicted peak value, decreasing the exposure of the image.

8. The method of claim 7, wherein according to the first formula, the predicted peak value is approximately an average of the luminance values minus a standard deviation of the luminance values times a skew of the luminance values.

9. The method of claim 7, wherein according to the second formula, the predicted peak value is approximately a total of the luminance values divided by the number of luminance values.

10. The method of claim 7, wherein the skew is calculated for a portion of the image that comprises pixels of interest.

11. The method of claim 10, wherein the pixels of interest are selected from a group consisting of identified movement in the image, identified skin portions in the image, identified portions of the image closest to the camera, pixels corresponding to a speaking participant, and identified warm portions of the image.

12. A method for exposure control for an image, comprising:

receiving video data for display on a display device in a videoconference, wherein the video data comprises a plurality of pixels corresponding to at least one image, wherein the at least one image comprises at least one human participant image;

determining a first plurality of pixels corresponding to a first portion of the image, wherein said determining the first plurality of pixels is performed based on identified skin portions of the at least one human participant image, wherein each of the plurality of pixels comprises a luminance value;

calculating a skew of the luminance values of the first plurality of pixels;

if the skew is above a predetermined value, calculating a predicted peak value of the luminance values according to a first formula;

if the skew is not above a predetermined value, calculating the predicted peak value of the luminance values according to a second formula;

if the predicted peak value is less than a predetermined low predicted peak value, increasing an exposure of the image; and

if the predicted peak value is more than a predetermined high predicted peak value, decreasing the exposure of the image.

13. A method for exposure control for an image, comprising:

receiving video data for display on a display device in a videoconference, wherein the video data comprises a plurality of pixels corresponding to at least one image, wherein the at least one image comprises at least one human participant image;

determining a first plurality of pixels corresponding to a first portion of the image, wherein said determining the first plurality of pixels is performed based on identified skin portions of the at least one human participant image, wherein each of the plurality of pixels comprises a luminance value;

determining a percentage of the first plurality of pixels versus the plurality of pixels comprising the image;

if the percentage is above a minimum skin percentage, calculating a skew of the luminance values of the first plurality of pixels;

if the percentage is not above a minimum skin percentage, calculating a skew of the luminance values for the plurality of pixels;

if the calculated skew is above a predetermined skew value, calculating a predicted peak value according to a first formula using the luminance values used to calculate the skew;

if the skew is not above a predetermined skew value, calculating the predicted peak value according to a second formula using the luminance values used to calculate the skew;

if the predicted peak value is less than a predetermined low predicted peak value, increasing an exposure of the image; and

if the predicted peak value is more than a predetermined high predicted peak value, decreasing the exposure of the image.

14. The method of claim 13, wherein according to the first formula, the predicted peak value is approximately an average of the luminance values minus a standard deviation of the luminance values times a skew of the luminance values.

15. The method of claim 13, wherein according to the second formula, the predicted peak value is approximately a total of the luminance values divided by the number of luminance values.

16. A system, comprising:

a camera; and

electronics coupled to the camera for performing a method comprising:

receiving video data for display on a display device in a videoconference, wherein the video data comprises a plurality of pixels corresponding to at least one image, wherein the at least one image comprises at least one human participant image;

determining a first plurality of pixels corresponding to a first portion of the image, wherein said determining the first plurality of pixels is performed based on identified skin portions of the at least one human participant image, wherein each of the plurality of pixels comprises a luminance value;

determining a percentage of the first plurality of pixels versus the plurality of pixels comprising the image;

if the percentage is above a minimum skin percentage, calculating a skew of the luminance values of the first plurality of pixels;

if the percentage is not above a minimum skin percentage, calculating a skew of the luminance values for the plurality of pixels;

if the calculated skew is above a predetermined skew value, calculating a predicted peak value according to a first formula using the luminance values used to calculate the skew;

if the skew is not above a predetermined skew value, calculating the predicted peak value according to a second formula using the luminance values used to calculate the skew;

if the predicted peak value is less than a predetermined low predicted peak value, increasing an exposure of the image; and

if the predicted peak value is more than a predetermined high predicted peak value, decreasing the exposure of the image.

17. The system of claim 16, wherein according to the first formula, the predicted peak value is approximately an average of the luminance values minus a standard deviation of the luminance values times a skew of the luminance values.

18. The system of claim 17, wherein according to the second formula, the predicted peak value is approximately a total of the luminance values divided by the number of luminance values.

19. The system of claim 17, wherein the minimum skin percentage varies based on one or more image characteristics.

20. The system of claim 19, wherein the image characteristic is previous image exposure.